Unsupervised Segmentation Method for Color Image Based on MRF
Yimin Hou, Xiangmin Lun, Wei Meng, Tao Liu, Xiaoli Sun · 2009
The paper proposes an unsupervised color image segmentation method based on Markov random field (MRF). The method involves intensity Euclidean distance and spatial position information of the pixels in the neighborhood potential function of MRF. Therefore, the traditional potential function of MRF segmentation method is improved. Transforms the segmentation to a maximum a posteriori (MAP) problem which is solved by the iterative conditional model (ICM). Uses the fuzzy C-means to initialize the classification in the rang of specified class number. The optimal class number was chosen according to minimum message length (MML) criterion to complete an unsupervised segmentation. In the experiments, synthetic and real images are used in the procedure and the results show that the proposed method is more effective than the classical methods.